Text to speech – AI voices, MP3 to save
Chat shortcut: /texttospeech
Paste text, choose a voice and listen: that's how you have texts of up to 20,000 characters per run turned from text to speech, and you save the result as MP3 for videos, podcasts or listening on the go.
With Account
From text to audio file in one step
Some texts are easier to understand when you hear them: your own drafts for proofreading, study material on the way to work, long articles while cooking. The tool reads pasted texts of up to 20,000 characters per run aloud with an AI voice, without you having to install anything.
The same function gives you voice recordings for projects: the voiceover for an explainer video, the announcement for a presentation, the intro for a podcast or an audio version of your website. You choose a voice, listen to the result and save it as MP3 if it fits.
The voices speak English, German and many other languages; the language is detected from the text. For a clean result, write numbers, abbreviations and proper names the way they should sound, structure the text with punctuation and paragraphs, and test-listen to critical passages.
How it works
- 1
Paste text
Copy the text into the field or type it in directly, up to 20,000 characters per run.
- 2
Choose a voice
Pick one of the voices, for example a calm, deep voice for explanations or a bright one for short announcements.
- 3
Read aloud
Click “Read aloud” and listen to the result right in the browser.
- 4
Save MP3
Download the audio file and use it in videos, presentations or on the go.
Typical uses
Proofreading by ear
Have your own text read aloud to hear clumsy sentences and typos you skip over when reading.
Learning on the go
Save summaries and notes as MP3 and listen while commuting or exercising.
Voiceover for videos
Generate a narrator voice for explainer videos, product clips and presentations without recording yourself.
Accessibility
Make texts audible for people who find reading hard, for example with impaired vision or dyslexia.
Good to know
- Voice selection depends on the active text-to-speech model: six voices with OpenAI, a list from the voice library with ElevenLabs; the language is detected from the text.
- Up to 20,000 characters per run; split longer texts into sections. Result as MP3 to listen to and download.
- Usage in tokens by character count and active model, available with an account.
- Emphasis and pauses can be influenced through punctuation and paragraphs; write out numbers and abbreviations so they are spoken as intended.
- The text is sent to the speech provider for generation; you get the MP3 back directly and can download it; with an account it is also kept in “My media”.
- In the chat you reach the tool with the shortcut /texttospeech.
Frequently asked questions
Which voices are there?
That depends on the active text-to-speech model: with OpenAI there are six voices with different character to choose from, from bright and friendly to deep and calm; with ElevenLabs a list from the voice library.
Do the voices sound natural or like a robot?
The voices come from current AI models and derive emphasis and pauses from the sentence structure. How natural the result sounds depends on voice, model and text; listen to it before you use it.
Does the voice pronounce other languages correctly?
The models detect the language from the text and speak English, German with umlauts and ß, and many other languages. For abbreviations, numbers and foreign words it helps to write them the way they should be spoken; check such passages when listening.
Can I control speed and emphasis?
There is no speed control. You influence it through the text: periods and paragraphs usually produce pauses, commas short breaths. How strong the effect is depends on the model.
How long can the text be?
Up to 20,000 characters per run. Split longer texts into sections and save each as its own MP3; join the files in an audio program if needed.
What does it cost?
Usage depends on the character count and the active model; a short paragraph costs correspondingly little. The tool is available with an account.
What happens to my text?
It is sent to OpenAI or ElevenLabs to generate the speech; what the provider stores is governed by its terms. You get the finished MP3 back directly; with an account it also stays in “My media” until you delete it there.